Result was reported too late to validate!?

Stefan
Stefan
Joined: 15 Nov 05
Posts: 52
Credit: 761198
RAC: 0
Topic 191273

I've gotten three results given 0 credit because the "result was reported too late to validate", even though the due date isn't till June...

I think something is messed up...

One of the results

Human Stupidity Is Infinite...

Erik
Erik
Joined: 14 Feb 06
Posts: 2815
Credit: 2645600
RAC: 0

Result was reported too late to validate!?


I'm also curious too about the quorum of 4. (Probably should've asked under the Problems and Bug Reports board but since it's already here...)

Ulrich Metzner
Ulrich Metzner
Joined: 22 Jan 05
Posts: 113
Credit: 963370
RAC: 0

RE: http://einstein.phys.uw

That one really looks suspicious :?
Maybe something gone wrong with the validators.
I also have lots of results, that i crunched and reported and they never made to the data base...
Just some recent (lots of them more, but they vanished with time) results:
http://einsteinathome.org/task/28887715
http://einsteinathome.org/task/28886386
http://einsteinathome.org/task/28869918
...just a excerpt from many... :/

Aloha, Uli

Stefan
Stefan
Joined: 15 Nov 05
Posts: 52
Credit: 761198
RAC: 0

RE: RE: http://einstein.p

Message 34138 in response to message 34137

Quote:
That one really looks suspicious :?
Maybe something gone wrong with the validators.
I also have lots of results, that i crunched and reported and they never made to the data base...
Just some recent (lots of them more, but they vanished with time) results:
http://einsteinathome.org/task/28887715
http://einsteinathome.org/task/28886386
http://einsteinathome.org/task/28869918
...just a excerpt from many... :/

I've got a ton of client error WU's too...and another one that received 0 credit..

Its really strange, there seem to be 2-3 pages worth of WU's labeled client error, I know there were 10 or so I lost the other day while reinstalling BOINC but 3 pages is 60 WU's...
Results

30704985 8680157 22 May 2006 22:36:04 UTC 5 Jun 2006 22:36:04 UTC Over Client error New 0.00 0.00 ---
30704979 8680155 22 May 2006 22:36:02 UTC 5 Jun 2006 22:36:02 UTC Over Client error New 0.00 0.00 ---
30704973 8680153 22 May 2006 22:36:00 UTC 5 Jun 2006 22:36:00 UTC Over Client error New 0.00 0.00 ---
30704970 8680152 22 May 2006 22:35:58 UTC 23 May 2006 17:40:09 UTC Over Success Done 3,338.32 8.57 45.56
30704813 8680102 22 May 2006 22:35:56 UTC 23 May 2006 17:13:10 UTC Over Success Done 3,468.13 8.90 0.00

Human Stupidity Is Infinite...

Bruce Allen
Bruce Allen
Moderator
Joined: 15 Oct 04
Posts: 1119
Credit: 172127663
RAC: 0

RE: I've gotten three

Quote:

I've gotten three results given 0 credit because the "result was reported too late to validate", even though the due date isn't till June...

I think something is messed up...

One of the results

You're absolutely right: something is messed up. This is a bug in the back-end BOINC server components, which I think was introduced on March 11th when David Anderson made some changes to the code. It has affected a couple of hundred results on Einstein@Home.

I've spent about 12 hours trying to track it down with no success yet, though I have inserted a bunch of debugging code into the scheduler which should make it easier to track down.

David Anderson checked and he has the same bug in SETI@Home, affecting about six hundred results. He's also looking to see if he can figure out what's wrong.

I am sure we can track this down and fix it. Thank you very much for pointing this out -- it seems to be the first scheduler/back-end bug that has shown up now in almost three months.

(By the way, it would be helpful to David and me to get a copy of your sched_request_einstein.phys.uwm.edu.xml file, and the corresponding sched_reply. Please don't post them here because they have your authenticator in them. Could you email them to me?)

Bruce

Director, Einstein@Home

Stefan
Stefan
Joined: 15 Nov 05
Posts: 52
Credit: 761198
RAC: 0

RE: RE: I've gotten three

Message 34140 in response to message 34139

Quote:
Quote:

I've gotten three results given 0 credit because the "result was reported too late to validate", even though the due date isn't till June...

I think something is messed up...

One of the results

You're absolutely right: something is messed up. This is a bug in the back-end BOINC server components, which I think was introduced on March 11th when David Anderson made some changes to the code. It has affected a couple of hundred results on Einstein@Home.

I've spent about 12 hours trying to track it down with no success yet, though I have inserted a bunch of debugging code into the scheduler which should make it easier to track down.

David Anderson checked and he has the same bug in SETI@Home, affecting about six hundred results. He's also looking to see if he can figure out what's wrong.

I am sure we can track this down and fix it. Thank you very much for pointing this out -- it seems to be the first scheduler/back-end bug that has shown up now in almost three months.

(By the way, it would be helpful to David and me to get a copy of your sched_request_einstein.phys.uwm.edu.xml file, and the corresponding sched_reply. Please don't post them here because they have your authenticator in them. Could you email them to me?)

Bruce

I just sent you the email with the file, and I should mention that I've had some strange things occuring with SETI as well. No zero credit yet but a ton of client error messages...

SETI errors

Human Stupidity Is Infinite...

Bruce Allen
Bruce Allen
Moderator
Joined: 15 Oct 04
Posts: 1119
Credit: 172127663
RAC: 0

Problem should now be

Problem should now be fixed.

Affected machine need to contact the scheduler twice before they start acting 'normal' again. You should cancel/abort any workunits currently in progress on the machine, as they have already been marked as 'errors' in the database. So CPU time spent on those in-progess results is wasted. Look at the 'my results' web page to see if that work has already been marked as 'client error'.

If your machine has this problem, your machine's sched_reply_einstein.phys.uwm.edu.xml file will contain
NUMBER
which has other text before it on the same line. This was caused by a bug in code used by Grid Republic.

Cheers,
Bruce

Director, Einstein@Home

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.